Full Terms & Conditions of access and use can be found athttps://www.tandfonline.com/action/journalInformation?journalCode=rjsp20

Journal of Sports Sciences

ISSN: 0264-0414 (Print) 1466-447X (Online) Journal homepage: https://www.tandfonline.com/loi/rjsp20

Development, construct validity and test–retestreliability of a field-based wheelchair mobilityperformance test for wheelchair basketball

Annemarie M. H. de Witte, Marco J. M. Hoozemans, Monique A. M. Berger,Rienk M. A. van der Slikke, Lucas H. V. van der Woude & Dirkjan (H. E. J)Veeger

To cite this article: Annemarie M. H. de Witte, Marco J. M. Hoozemans, Monique A. M. Berger,Rienk M. A. van der Slikke, Lucas H. V. van der Woude & Dirkjan (H. E. J) Veeger (2018)Development, construct validity and test–retest reliability of a field-based wheelchair mobilityperformance test for wheelchair basketball, Journal of Sports Sciences, 36:1, 23-32, DOI:10.1080/02640414.2016.1276613

To link to this article: https://doi.org/10.1080/02640414.2016.1276613

View supplementary material Published online: 16 Jan 2017.

Submit your article to this journal Article views: 369

View Crossmark data Citing articles: 6 View citing articles

Development, construct validity and test–retest reliability of a field-basedwheelchair mobility performance test for wheelchair basketballAnnemarie M. H. de Witte a,b, Marco J. M. Hoozemansb, Monique A. M. Bergera, Rienk M. A. van der Slikkea,c,Lucas H. V. van der Wouded,e, and Dirkjan (H. E. J) Veegerb,c

aFaculty Health, Nutrition & Sport, The Hague University of Applied Sciences, The Hague, the Netherlands; bDepartment of Human Movement Sciences,Vrije Universiteit Amsterdam, MOVE Research Institute, Amsterdam, the Netherlands; cDepartment of Biomechanical Engineering, University ofTechnology, Delft, the Netherlands; dCentre for Human Movement Sciences, University Medical Centre Groningen, University of Groningen, Groningen,the Netherlands; eCentre for Rehabilitation, University Medical Centre Groningen, University of Groningen, Groningen, the Netherlands

ABSTRACTThe aim of this study was to develop and describe a wheelchair mobility performance test in wheelchairbasketball and to assess its construct validity and reliability. To mimic mobility performance of wheel-chair basketball matches in a standardised manner, a test was designed based on observation ofwheelchair basketball matches and expert judgement.

Forty-six players performed the test to determine its validity and 23 players performed the test twicefor reliability. Independent-samples t-tests were used to assess whether the times needed to completethe test were different for classifications, playing standards and sex. Intraclass correlation coefficients(ICC) were calculated to quantify reliability of performance times.

Males performed better than females (P < 0.001, effect size [ES] = −1.26) and international menperformed better than national men (P < 0.001, ES = −1.62). Performance time of low (≤2.5) and high(≥3.0) classification players was borderline not significant with a moderate ES (P = 0.06, ES = 0.58). Thereliability was excellent for overall performance time (ICC = 0.95).

These results show that the test can be used as a standardised mobility performance test to validlyand reliably assess the capacity in mobility performance of elite wheelchair basketball athletes.Furthermore, the described methodology of development is recommended for use in other sports todevelop sport-specific tests.

ARTICLE HISTORYAccepted 18 December 2016

KEYWORDSSkill performance;wheelchairs; taskperformance and analysis;paralympic

Introduction

In wheelchair court sports, the player, the wheelchair and theenvironment determine performance. All the activities an ath-lete does (or can do) with a wheelchair, the wheelchair–athleteactivities, can be defined as mobility performance. Key deter-minants of mobility performance are the abilities of the athleteto accelerate, sprint, brake and turn with the wheelchair (deWitte, Hoozemans, Berger, Veeger, & van der Woude, 2016;Mason, Porcellato, van der Woude, & Goosey-Tolfrey, 2010).The actual mobility performance in wheelchair court sportsshould be assessed during a match, preferably by systematic(video) observation combined with the use of (inertial) sensors(Bloxham, Bell, Bhambhani, & Steadward, 2001; de Witte et al.,2016; Rhodes, Mason, Perrat, Smith, & Goosey-Tolfrey, 2014;Van der Slikke, Berger, Bregman, & Veeger, 2015). These obser-vations and measurements during wheelchair basketball resultin, for example, findings that players move across the fieldwith light or no arm strokes for 24% (standard deviation [SD]7) of the time (Bloxham et al., 2001) and that national standardplayers drive relatively more forward, while international stan-dard players perform more rotational movements during amatch (de Witte et al., 2016). Assessing mobility performance

is a fundamental requirement for trainers and coaches to, forexample, develop training schemes, discuss and improve theathlete’s level of performance, detect strength and weak-nesses of mobility performance and develop optimal wheel-chair configurations. The use of systematic observation and/orsensor technology during matches can thus provide usefulinformation about mobility performance. However, systematicobservation is very time-consuming, and results of both meth-ods are influenced by the continuously changing environmentwhen participating in a match of wheelchair basketball. Eachmatch has unique circumstances depending on, for example,the opponent, injuries or team composition.

In order to repeatedly monitor athletes’ mobility perfor-mance, athlete performance on a standardised field-basedtest is assigned to be informative and helpful (Goosey-Tolfrey & Leicht, 2013; Vanlandewijck, Daly, & Theisen, 1999).Currently, there is no generally accepted validated mobilityperformance test available for wheelchair court sports in gen-eral and for wheelchair basketball specifically. To assess andmonitor mobility performance in a controllable setting, themobility performance during a match must be simulated. Asimulation or test that is based on field activities – i.e., the

CONTACT Annemarie M. H. de Witte a.m.h.dewitte@hhs.nl Faculty Health, Nutrition and Sport, The Hague University of Applied Sciences, Laan van Poot363, 2566DA, The Hague, the Netherlands.

Supplemental data for this article can be accessed here.

JOURNAL OF SPORTS SCIENCES, 2018VOL. 36, NO. 1, 23–32https://doi.org/10.1080/02640414.2016.1276613

© 2017 Informa UK Limited, trading as Taylor & Francis Group

match – will result in meaningful information for coaches,players and (embedded) scientists. Field-based tests are gen-erally acknowledged as a feasible way to get an indication ofthe performance standard of athletes (de Groot, Balvers,Kouwenhoven, & Janssen, 2012). Field-based tests exist forwheelchair court sports, but they assess mainly other aspectsof performance, such as game performance (ball skills) andathlete performance (e.g., maximal heart rate or oxygen con-sumption) and only some parts of mobility performance(Barfield & Malone, 2012; Byrnes & Hedrick, 1994; de Grootet al., 2012; de Groot, Valent, Fickert, Pluim, & Houdijk, 2016;Gil et al., 2015; Granados et al., 2015; Yilla & Sherrill, 1998).

Extensive systematic observation and analyses of mobilityperformance during wheelchair basketball matches haverecently been done for wheelchair basketball (de Witte et al.,2016; Van der Slikke et al., 2016). These data were used todevelop a standardised and worldwide-accepted wheelchairmobility performance (WMP) test. Feasibility is a preconditionin the development process, and the test should be easy totake without advanced equipment. To further ensure a highexternal validity, the test should be performed by wheelchairbasketball players in their own sports wheelchair and on aregular wheelchair basketball court. Furthermore, the testshould discriminate between different categories of athletes(e.g., sex and playing standard), which is known from theliterature that they differ in mobility performance (de Witteet al., 2016; Gomez, Perez, Molik, Szyman, & Sampaio, 2014;Van der Slikke, Berger, Bregman, & Veeger, 2015; Van derSlikke et al., 2016; Vanlandewijck, Daly, Spaepen, Theisen, &Pétré, 1999). Besides valid results, the test should give reliabledata to monitor the actual capacity in mobility performance ofathletes.

In this context, the goals of the present study were (1) todescribe the development of a field-based wheelchair test thatassesses mobility performance capacity and which closelymimics the wheelchair mobility skills required in real wheel-chair basketball matches, (2) to define the developed field-based test and (3) to assess the construct validity and test–retest reliability of the newly developed field-based WMP testfor wheelchair basketball.

Methods

Test development

The development process had a stepwise character: (1) exam-ine match mobility performance, (2) determine practical testrequirements and (3) organise expert meetings to verify thetest design.

To examine mobility performance in matches, coacheswere interviewed to describe and define wheelchair–athleteactivities during wheelchair basketball. The wheelchair activ-ities were assessed by systematic observation of video footageof matches (de Witte et al., 2016). Four matches at nationalplaying standard and five matches at international playingstandard were recorded. In total, 56 male wheelchair basket-ball players were analysed during an entire match. Time–motion analysis was used for determining the frequency andduration of these athlete and wheelchair activities (de Witteet al., 2016). Based on the results, wheelchair basketball mobi-lity performance was defined in various dominant game-related wheelchair activities (Table 1). In order to make atranslation from match data to test design, the output wasorganised into three main categories: separate activities, com-bined activities and activities with ball possession. For each ofthese categories, the most common wheelchair–athlete activ-ities and distances were determined with inertial sensors (Vander Slikke, Berger, Bregman, Lagerberg, & Veeger, 2016).

In addition, practical test requirements were formulated forthe WMP test based on interviews with coaches and experts:(1) The WMP test should be easy to use without advancedequipment; (2) The WMP test should take place in a realisticenvironment common to wheelchair basketball, e.g., athletesperformed the test in their own sports wheelchairs and on aregular wheelchair basketball court; and (3) Fatigue should notbe a limiting factor for performance. The observed activitiesand the requirements were used to draft the first test setup.

An expert meeting with coaches, players and researcherswas organised to discuss the first version of the WMP test toincrease its content validity, after which “specific skills” wereadded as a fourth main group. The four main groups con-tained a total of 15 different wheelchair–athlete activities(Table 2). Based on these data, a final version of the WMPtest was developed which is described in the results section.The development process took place between March 2014and March 2015.

Construct validity and test–retest reliability

To evaluate the construct validity and reliability of the newlydeveloped WMP test, experienced wheelchair basketballplayers were included in different field-based standardisedexperimental sessions.

ParticipantsFor the validity study, 46 players – competing at differentplaying standards – were included, and for the reliability

Table 1. Overview of the relative duration (±SD) as a percentage of wheelchair–athlete activities based on video analysis of 56 male wheelchair basketball athletesplaying at national and international playing standard (de Witte et al., 2016).

Wheelchair activities Outcome video analysis relative duration % (±SD) Relative duration during ball possession % (±SD) Outcome inertial sensors

Standing still 19 (6) 26 (16) –Driving forward 45 (6) 42 (12) Most common: 3 m

Maximal: 12 mDriving backward 2 (1) 1 (1) –Rotate 29 (8) 28 (12) Most common: radius 1.5–2.5 mBrake 3 (2) 2 (2) –

The data are complemented with information from data of inertial sensors based on 29 wheelchair basketball players (Van der Slikke et al., 2016).

24 A. M. H. DE WITTE ET AL.

study, 23 players – competing at a national playing standard(Dutch first division competition) – participated. In the validitygroup, a distinction was made between men and womencompeting at an international standard and players compet-ing at a national standard, and a distinction was madebetween low classification (≤2.5 points) and high classification(≥3.0 points) players. The International Wheelchair BasketballFederation uses a classification system based on the players’functional potential to execute fundamental basketball move-ments (International Wheelchair Basketball Federation, 2014).All players are scaled from 1 (minimal functional potential) to4.5 points (maximal functional potential) on an ordinal func-tional level scale. The characteristics (classification, basketballexperience and age) of the validity and reliability study groupsare shown in Table 3. Players were informed about the proce-dures before giving their written informed consent. This studywas approved by the Ethical Committee of the Department ofHuman Movement Sciences, Vrije Universiteit Amsterdam, theNetherlands.

ProcedurePrior to all tests, procedures were explained, and the testprotocol was demonstrated using a video shown to all parti-cipants. Players were asked to refrain from smoking and drink-ing caffeine or alcohol at least 2 h prior to the WMP test.Before performing the WMP test, players carried out a self-selected warm-up. All players performed the WMP test in theirown sports wheelchairs, with their own configurations andtires were inflated to 7 bar.

Participants of the validity study performed the WMP test onceon the same synthetic soft-top basketball court. Participants weremeasured while being involved in training sessions and in theEuro Cup 4 tournament (April 2015, the Netherlands).

Participants of the test–retest reliability study performedthe same test twice. Participants were tested during theirtraining sessions, on the basketball courts where the teamstrained, on two separate days at the same time of the day,with 1 week in between (October/November 2015).

Data acquisition and analysesThe WMP test simulated the 15 most common wheelchair–athlete activities during wheelchair basketball (Table 2). All thestandardised activities were carried out in succession, sepa-rated by standardised rest periods to avoid fatigue. Two high-definition video cameras (CASIO EX-FH100, 1280*720,20–240 mm) were placed at the side of the test. Each camera

was focused on one half of the basketball court with a smalloverlap between the videos. The outcome of the WMP testwas time (s), which was manually recorded from video analysis(Kinovea 0.8.15, France). These analyses resulted in 16 perfor-mance time values, one for each of the 15 wheelchair–athleteactivities (time activity no. 1–15) and the overall performancetime, which is the sum of the performance times of the 15separate activities.

Statistical analysesAll statistical analyses were performed using IBM SPSS statis-tics version 22 (IBM Corporation, Armonk, NY, USA).Descriptive statistics for the time activities no. 1–15 and theoverall performance time were presented as mean ± SD. Theassumptions of normality were checked with the Shapiro–Wilktest, as well as z-values of the skewness and kurtosis. Also,histograms, boxplots and q-q plots of the data were visuallyinspected. The assumption of normality was not violated.

Construct validityTo determine the construct validity of the WMP test, threehypotheses were formulated and tested. Hypothesis (1):Players with a high classification (≥3.0 points) are expected toperform better than players with a low classification (≤2.5points) (Van Der Slikke, Berger, Bregman, & Veeger, 2015;Vanlandewijck, Daly, & Theisen, 1999). Hypothesis (2): Playersplaying at an international standard are expected to performbetter than players at a national standard (de Witte et al., 2016;Van Der Slikke et al., 2015). Hypothesis (3): Men are expected toperform better than women because of sex differences in upperbody strength and trunk stability as key determinants of mobi-lity performance (Gómez et al., 2014).

To assess potential differences in the 16 performance timeoutcomes between classification categories, playing standardsand sex, independent samples t-tests were used. The means ±SDs were completed with mean differences, 95% confidenceintervals of the difference and P-values. Differences withP-values <0.05 were considered statistically significant. In addi-tion, Cohen’s d effect sizes (ES) were calculated for maineffects as outlined by Cohen (1992). The (absolute) magnitudeof the ES was classified as large (≥0.80), moderate (0.50–0.79)or small (<0.50) (Cohen, 1988).

Test–retest reliabilityTest–retest reliability of the 16 time performance outcomeswas evaluated with intraclass correlation coefficients (ICC(3,1)),

Table 2. Setup test protocol based on observed wheelchair–athlete activities and distances (for the total test protocol, see Supplementary material I).

Main group Activity Distance Direction

Separate activities Driving forward 12 m –Rotation Radius 1.9 m (total circumference of 12 m) Clockwise/counterclockwiseRotation on the spot Clockwise/counterclockwise

Combined activities Driving forward with two stops 3, 3 and 6 m = 12 m –Rotation with two stops 90° (3 m), 90° (3 m), 180° (6 m) = 12 m Clockwise/counterclockwiseRotation on the spot with stop 90°, 90° Clockwise/counterclockwiseCombined activities –

Specific skills Tik-tak box –Activities with ball possession Driving forward 12 m –

Rotation Radius 1.9 m (total circumference of 12 m) Clockwise/counterclockwise

JOURNAL OF SPORTS SCIENCES 25

Table3.

General

characteristicsof

theparticipantsinclud

edin

theconstructvalidity

(n=46)andtest–retestreliability(n

=23)analyses

forclassification1–4.5.

Classification

nExperiencein

years(±SD

)

Agein

years

(±SD

)Mean(±SD

)andrang

eof

wheel

size

(cm)

Mean(±SD

)andrang

eof

elbo

wangle

with

hand

onthetopof

therim

(°)

Mean(±SD

)andrang

eof

wheel

camber(°)

Men

playingat

internationalstand

ard

(n)

Wom

enplayingat

internationalstand

ard

(n)

Men

playingat

natio

nalstand

ard

(n)

Validity

stud

y1–1.5

87.2(4.8)

28.3

(7.1)

62.0

(2.4)

100(11)

17(1)

33

258–64

86–122

16–19

2–2.5

1112.9

(6.9)

28.9

(9.3)

62.8

(2.6)

117(18)

17(1)

63

259–68

77–135

15–19

3–3.5

89.1(3.3)

26.7

(10.0)

64.4

(1.1)

128(18)

18(1)

53

–64–67

100–162

17–21

4–4.5

198.4(5.2)

24.7

(8.3)

64.5

(2.0)

136(18)

18(1)

74

861–68

99–168

15–21

Reliability

stud

y1–1.5

24.0(0.7)

21.0

(4.2)

61.5

(3.5)

87(1)

17(1)

––

259–64

86–88

16–17

2–2.5

19.0

21.0

61.0

110

17–

–1

3–3.5

56.4(1.9)

16.8

(5.1)

60.4

(2.9)

104(24)

18(2)

––

558–64

81–136

15–20

4–4.5

156.5(6.4)

22.8

(10.8)

63.4

(2.5)

129(16)

18(1)

––

1556–67

99–151

15–20

26 A. M. H. DE WITTE ET AL.

standard error of measurement (SEM) and limits of agreement(LoA). ICC(3,1) is a two-way mixed single measure of absoluteagreement (Shrout & Fleiss, 1979). ICC scores ≥0.70 are indi-cated as satisfactory, values ≥0.75 are considered as good andvalues ≥0.90 are categorised as excellent reliability (Atkinson &Nevill, 1998). The SEM for agreement was calculated withEquation (1).

SEM agreement ¼ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiVaro þ Varresidual

p(1)

Variance components were obtained from variance compo-nent analyses, and two components were estimated, varianceattributable to observers (Varo) and residual error (Varresidual).

The Bland–Altman method was used to examine the differ-ences between the WMP test and retest for the whole group,including the calculation of the mean difference between thetest and retest, the SD of the difference, and the 95% LoA (Bland& Altman, 1986). The LoA95 was calculated with Equation (2).

LoA95 ¼ meandifference� 1:96 � SDdifference (2)

The differences for the overall performance times were visualisedin a Bland–Altman plot, where the individual differences betweenthe test and retest are plotted against the mean of the test andretest.

Results

Design of the WMP test

The final version of the WMP test for wheelchair basketballconsisted of 15 activities with a standardised period of restbetween the activities. The WMP test is divided into four maingroups. Group (1): Separate activities containing a 12 m sprint,a rotation with a curve (circumference) of 12 m (clockwise/

counterclockwise) and a turn on the spot (clockwise/counter-clockwise); Group (2): Combined activities containing the sameactivities as group 1, combined with starts and stops inbetween; Group (3): Specific skills consisting of a tik-tak box,which means performance of short movements forward andbackward alternated with collisions against a stationary object.Group (4): A 12 m sprint and rotation (clockwise/counterclock-wise) with a curve (circumference) of 12 m performed with ballpossession (dribble) (for the total WMP test protocol and thesequence of the activities, see Supplementary material I).

Construct validity and test–retest reliability

Time scores of the tik-tak box (activity no. 1) of the WMP testwere not included in both the reliability and the constructvalidity study. The start and stop times of this activity were notclearly visible at the video analysis, and because of this, thedata are not presented and included.

Construct validityTo determine the construct validity of the WMP test, threehypotheses were formulated and tested.

Hypothesis (1): Players with a high classification are expectedto perform better than players with a low classification. Theoverall performance time was borderline non-significantbetween high and low classifications (P = 0.06, ES = 0.58),but the magnitude of the ES can be interpreted as moderate(Table 4). For time scores on the individual activities, theclassification analyses showed significant differences for driv-ing forward movements and turn on the spots, in which high-classification players performed the activities faster than low-classification players. Significant differences between high andlow classifications were observed for the 12 m sprint (meandifference = 0.32 s; ES = 0.92) and for the 3-3-6 m sprint (mean

Table 4. Mean (±SD) performance times (s) for each activity and overall performance time (s) of the wheelchair mobility performance test for classification(classification ≤2.5 points and classification >2.5 points) complemented with the mean difference between the classification groups, 95% confidence intervals of thedifferences and Cohen’s d effect sizes.

Classification ≤2.5points (n = 19)

Classification >2.5points (n = 27)

95%confidence

interval of thedifference

Mean (±SD) Mean (±SD)Mean

differenceStandard errordifference Lower Upper P-values

Effectsize

Activity 2 180° turn on the spot (left) 0.93 (0.09) 0.84 (0.08) 0.09 0.02 0.04 0.14 0.00a 1.04Activity 3 12 m sprint 5.12 (0.42) 4.80 (0.28) 0.32 0.10 0.11 0.53 0.00a 0.92Activity 4 12 m rotation (right) 5.97 (0.41) 5.90 (0.40) 0.07 0.12 −0.17 0.31 0.57 0.17Activity 5 12 m rotation (left) 5.95 (0.47) 5.89 (0.39) 0.06 0.13 −0.19 0.32 0.62 0.15Activity 6 180° turn on the spot

(right)0.95 (0.13) 0.89 (0.12) 0.06 0.04 −0.01 0.14 0.10 0.50

Activity 7 3-3-6 m sprint 7.19 (0.77) 6.64 (0.61) 0.55 0.20 0.14 0.96 0.01a 0.81Activity 8 3-3-6 m rotation (left) 7.66 (0.84) 7.33 (0.61) 0.33 0.21 −0.10 0.76 0.13 0.47Activity 9 3-3-6 m rotation (right) 7.58 (0.80) 7.23 (0.61) 0.36 0.21 −0.06 0.78 0.09 0.51Activity 10 90°–90° turn on the spot

with stop (left)1.54 (0.19) 1.38 (0.17) 0.16 0.05 0.05 0.27 0.01a 0.87

Activity 11 12 m dribble 6.03 (0.70) 5.80 (0.68) 0.24 0.21 −0.18 0.65 0.26 0.34Activity 12 12 m rotation dribble

(right)7.38 (0.91) 7.17 (0.87) 0.22 0.26 −0.31 0.75 0.41 0.25

Activity 13 12 m rotation dribble (left) 7.42 (0.97) 7.27 (0.68) 0.15 0.24 −0.34 0.64 0.54 0.19Activity 14 90°–90° turn on the spot

with stop (right)1.41 (0.17) 1.31 (0.15) 0.10 0.05 0.00 0.19 0.05a 0.61

Activity 15 Combination 13.95 (0.95) 13.42 (0.67) 0.53 0.24 0.04 1.02 0.03a 0.67Overall performance time (sum activities 2–15) 79.25 (6.56) 75.95 (4.97) 3.30 1.72 −0.17 6.77 0.06 0.58aSignificant effect of classification (P < 0.05).

JOURNAL OF SPORTS SCIENCES 27

difference = 0.55 s; ES = 0.81). However, for nearly all activitiesrelated to rotation (7 out of 10), there was no differencebetween classification categories.

Hypothesis (2): Players playing at an international standardare expected to perform better than players at a national stan-dard. The WMP test showed a significant difference for playingstandard for the overall performance time (P < 0.001,ES = −1.62). International men performed the WMP test onaverage 8.11 s faster than the national men (Table 5). TheWMP test showed a significant difference between interna-tional men and national men for 13 of the 15 outcomes andshowed that international men were faster on all the activities(moderate/large ES: 0.81–1.72). The WMP test showed nodifferences for three of the four activities that measured turnon the spot (no. 2, 6 and 10) (moderate/small ES: 0.71–0.22).

Hypothesis (3): Men are expected to perform better thanwomen,both competing at the same playing standard. There was a sig-nificant difference between men and women on the overallperformance time (P < 0.001, ES = −1.26). International menperformed the WMP test faster than international women(Table 6). In addition, the WMP test showed differences betweeninternational men and international women on all activities withthe exception of the activities that measured turn on the spotand 12 m dribble. A striking detail is that international womenperformed the rotation on the spot activities almost as fast as theinternational men (small ES: 0.02–0.44).

Test–retest reliabilityThe test–retest reliability analyses results are summarised inTable 7. The ICC value for the overall performance time was

Table 6. Mean (±SD) performance times (s) for each activity and overall performance time (s) of the wheelchair mobility performance test for differences in sex(international men and international women) complemented with the mean difference between the sex groups, 95% confidence intervals of the differences, andCohen’s d effect sizes.

International men(n = 21)

International women(n = 13)

Meandifference

Standard errordifference

95%confidence

interval of thedifference

P-valuesEffectsizeMean (±SD) Mean (±SD) Lower Upper

Activity 2 180° turn on the spot (left) 0.87 (0.09) 0.89 (0.07) −0.02 0.03 −0.08 0.04 0.58 −0.20Activity 3 12 m sprint 4.76 (0.34) 5.04 (0.27) −0.28 0.11 −0.50 −0.05 0.02a −0.90Activity 4 12 m rotation (right) 5.72 (0.42) 6.07 (0.21) −0.35 0.12 −0.60 −0.09 0.01a −0.98Activity 5 12 m rotation (left) 5.67 (0.38) 6.07 (0.29) −0.40 0.12 −0.65 −0.15 0.00a −1.15Activity 6 180° turn on the spot (right) 0.90 (0.15) 0.90 (0.07) 0.00 0.04 −0.09 0.09 0.95 0.02Activity 7 3–3-6 m sprint 6.57 (0.75) 7.06 (0.52) −0.49 0.24 −0.97 −0.01 0.05a −0.73Activity 8 3–3-6 m rotation (left) 7.01 (0.71) 7.83 (0.45) −0.81 0.22 −1.27 −0.36 0.00a −1.30Activity 9 3–3-6 m rotation (right) 6.91 (0.56) 7.65 (0.56) −0.74 0.20 −1.14 −0.34 0.00a −1.33Activity 10 90°–90° turn on the spot with

stop (left)1.41 (0.21) 1.40 (0.14) 0.01 0.07 −0.14 0.15 0.93 0.03

Activity 11 12 m dribble 5.66 (0.63) 5.95 (0.70) −0.30 0.23 −0.77 0.17 0.21 −0.45Activity 12 12 m rotation dribble (right) 6.77 (0.69) 7.44 (0.84) −0.67 0.26 −1.20 −0.13 0.02a −0.89Activity 13 12 m rotation dribble (left) 6.88 (0.73) 7.47 (0.51) −0.58 0.23 −1.06 −0.11 0.02a −0.89Activity 14 90°–90° turn on the spot with

stop (right)1.28 (0.15) 1.34 (0.10) −0.06 0.05 −0.15 0.04 0.22 −0.44

Activity 15 Combination 13.15 (0.70) 13.88 (0.55) −0.73 0.23 −1.20 −0.26 0.00a −1.12Overall performance time (sum activities 2–15) 73.44 (4.95) 79.21 (3.88) −5.76 1.63 −9.08 −2.44 0.00a −1.26aSignificant effect of sex (P < 0.05).

Table 5. Mean (±SD) performance times (s) for each activity and overall performance time (s) of the wheelchair mobility performance test for differences in playingstandard (international men and national men) complemented with the mean difference between the (international) groups, 95% confidence intervals of thedifferences, and Cohen’s d effect sizes.

International men(n = 21)

National men(n = 12)

Meandifference

Standard errordifference

95% confidenceinterval of thedifference

P-valuesEffectsizeMean (±SD) Mean (±SD) Lower Upper

Activity 2 180° Turn on the spot (left) 0.87 (0.09) 0.89 (0.12) −0.02 0.04 −0.10 0.05 0.54 −0.22Activity 3 12 m sprint 4.76 (0.34) 5.08 (0.45) −0.32 0.14 −0.60 −0.03 0.03a −0.84Activity 4 12 m rotation (right) 5.72 (0.42) 6.16 (0.37) −0.43 0.15 −0.73 −0.14 0.01a −1.08Activity 5 12 m rotation (left) 5.67 (0.38) 6.17 (0.38) −0.51 0.14 −0.79 −0.23 0.00a −1.33Activity 6 180° Turn on the spot (right) 0.90 (0.15) 0.95 (0.15) −0.05 0.05 −0.16 0.06 0.38 −0.32Activity 7 3–3-6 m sprint 6.57 (0.75) 7.17 (0.73) −0.60 0.27 −1.15 −0.06 0.03a −0.81Activity 8 3–3-6 m rotation (left) 7.01 (0.71) 7.88 (0.52) −0.86 0.24 −1.34 −0.38 0.00a −1.32Activity 9 3–3-6 m rotation (right) 6.91 (0.56) 7.89 (0.60) −0.99 0.21 −1.41 −0.56 0.00a −1.72Activity 10 90°- 90° turn on the spot with

stop (left)1.41 (0.21) 1.55 (0.18) −0.14 0.07 −0.29 0.01 0.06 −0.71

Activity 11 12 m dribble 5.66 (0.63) 6.25 (0.67) −0.59 0.23 −1.07 −0.12 0.02a −0.92Activity 12 12 m rotation dribble (right) 6.77 (0.69) 7.91 (0.77) −1.13 0.26 −1.67 −0.60 0.00a −1.57Activity 13 12 m rotation dribble (left) 6.88 (0.73) 7.99 (0.72) −1.10 0.26 −1.64 −0.57 0.00a −1.52Activity 14 90°–90° turn on the spot with

stop (right)1.28 (0.15) 1.49 (0.17) −0.21 0.06 −0.32 −0.09 0.00a −1.34

Activity 15 Combination 13.15 (0.70) 14.17 (0.86) −1.02 0.28 −1.59 −0.45 0.00a −1.34Overall performance time (sum activities 2–15) 73.44 (4.95) 81.55 (5.08) −8.11 1.83 −11.84 −4.37 0.00a −1.62aSignificant effect of playing standard (P < 0.05).

28 A. M. H. DE WITTE ET AL.

excellent (ICC = 0.95). The LoA95 show that an improvement of4.20 s (5.1%) can be detected as a real improvement on theWMP test. The Bland–Altman plot for test–retest agreement ofthe overall performance time is shown in Figure 1. The meandifference between the WMP test and retest for the overallperformance time was 0.57 s (±2.14). The variability of thedifferences between the two measurements seems to be con-stant over the range of the (mean) performance time scores.The ICC values for the individual activities ranged from 0.25 forthe 180° turn on the spot (left) (no. 2) to 0.92 for the combina-tion (no. 15). The four activities that measured turn on the spot(no. 2, 6, 10 and 14) show a low reliability (ICC ≤0.62), while theLoA95 for these activities were high (at least 0.3 s, 22.0%).

Discussion

This study describes the development of a new field-basedWMP test to assess the capacity of mobility performance andits construct validity and test–retest reliability. To examine theconstruct validity, we hypothesised that classification, playingstandard and sex will influence the performance on the test.

The construct validity tests showed that the WMP test distin-guishes sex and playing standards, but did not show differ-ences between low and high classifications on the overallperformance time. The test–retest reliability for the overallperformance time was excellent, and an improvement of4.2 s (5.1%) can be detected relative to the overall perfor-mance time. However, the reliability for the activities relatedwith rotation on the spot and the 12 m sprint is low.

Test development

The WMP test which is introduced in this article is a simulationof mobility performance during matches specific to wheelchairbasketball. The WMP test can easily be used by trainers, coa-ches and scientists to gain insight into the capacity of mobilityperformance of players. The developed WMP test meets therequirements which have been reported in previous studies ofwheelchair court sports (Goosey-Tolfrey & Leicht, 2013; Mason,van der Woude, & Goosey-Tolfrey, 2013; Vanlandewijck,Theisen, & Daly, 2001). The WMP test is based on the mostcommon aspects of mobility performance, the players are

Table 7. Descriptive values of 23 national male wheelchair basketball players (mean (s) ±SD) and mean differences for the test–retest complemented with reliabilitystatistics (s): ICC(3,1) absolute agreement, 95% confidence interval of the ICC agreement, SEM and 95% limits of agreement.

Test 1 Test 2 Meandifference (±SD)

ICCagreement

95%confidence

interval of theICC agreement SEM

agreementLimits ofagreementMean (±SD) Mean (±SD) Lower Upper

Test 2 180° turn on the spot (left) 0.90 (0.15) 0.90 (0.10) 0.00 (0.15) 0.25 −0.19 0.60 0.10 0.30Test 3 12 m sprint 5.02 (0.36) 5.13 (0.42) −0.10 (0.34) 0.62 0.29 0.82 0.24 0.66Test 4 12 m rotation (right) 6.33 (0.56) 6.33 (0.49) 0.00 (0.23) 0.91 0.80 0.96 0.16 0.45Test 5 12 m rotation (left) 6.33 (0.54) 6.40 (0.56) −0.08 (0.31) 0.84 0.66 0.93 0.22 0.61Test 6 180° turn on the spot (right) 0.93 (0.16) 0.90 (0.13) 0.03 (0.14) 0.55 0.20 0.78 0.10 0.26Test 7 3-3-6 m sprint 7.11 (0.61) 6.98 (0.62) 0.14 (0.38) 0.80 0.58 0.91 0.28 0.75Test 8 3-3-6 m rotation (left) 8.05 (0.74) 7.92 (0.81) 0.13 (0.36) 0.88 0.74 0.95 0.26 0.70Test 9 3-3-6 m rotation (right) 8.06 (0.88) 7.82 (0.72) 0.24 (0.48) 0.79 0.53 0.91 0.37 0.94Test 10 90°–90° turn on the spot with

stop (left)1.49 (0.26) 1.40 (0.18) 0.09 (0.19) 0.62 0.28 0.82 0.14 0.37

Test 11 12 m dribble 6.23 (0.68) 6.19 (0.60) 0.04 (0.45) 0.76 0.51 0.89 0.31 0.88Test 12 12 m rotation dribble (right) 8.29 (1.31) 8.34 (1.20) −0.05 (0.81) 0.80 0.59 0.91 0.56 1.58Test 13 12 m rotation dribble (left) 8.30 (1.06) 8.24 (1.04) 0.06 (0.74) 0.76 0.52 0.89 0.51 1.44Test 14 90°–90° turn on the spot with

stop (right)1.40 (0.20) 1.36 (0.16) 0.04 (0.16) 0.62 0.30 0.82 0.11 0.31

Test 15 Combination 14.44 (1.30) 14.41 (1.13) 0.04 (0.49) 0.92 0.83 0.97 0.34 0.96Overall performance time (sum activities 2–15) 82.88 (7.22) 82.31 (6.41) 0.57 (2.14) 0.95 0.89 0.98 0.98 4.20

5

4

3

2

1

0

-1

-2

-3

-4

6

-5

Mean of performance time test and re-test

Overall performance time

Dif

fe

re

nce

be

tw

ee

n t

est a

nd

re

-te

st

60 65 70 75 80 85 90 95 100

Mean + 1.96SD

4.76

Mean - 1.96SD

-3.63

Mean

0.57

Figure 1. Bland–Altman graph for overall performance time on the wheelchair mobility performance test (n = 23). The solid line represents the mean differencebetween the test and retest. The dashed lines represent the 95% limits of agreement for the performance times (mean difference ± 1.96SD).

JOURNAL OF SPORTS SCIENCES 29

tested in their natural environment and they are tested withtheir own wheelchair configuration. However, mobility perfor-mance may change when essential aspects of the sportchange, e.g., changes in the basketball rulings or wheelchairregulations. In the case of such changes, the mobility perfor-mance needs to be redefined.

Construct validity

Players with a high classification (≥3.0 points) are expected toperform better than players with a low classification (≤2.5points) (Van der Slikke, Berger, Bregman, & Veeger, 2015;Vanlandewijck, Daly, & Theisen, 1999). The key determinantsof the classification system are the ability to have active stabilityand rotation possibilities of the trunk (International WheelchairBasketball Federation, 2014). Previous research shows that trunkimpairment had impact on wheelchair propulsion, especially inaccelerating from standstill (Chow et al., 2009; Vanlandewijcket al., 2001). The overall performance time of the WMP testshowed a borderline non-significant difference (P = 0.06) and amoderate ES in capacity of mobility performance between lowand high classifications. There were significant differencesbetween classification levels on the separate activities relatedto driving forward movements (no. 3, 7 and 15). In contrast,almost all activities related to rotational movements of thewheelchair showed no significant differences, which couldmean that classification (trunk impairment) has less influenceon rotational movements. Furthermore, the used cutoff pointfor dichotomising classification in this study is debatable. Otherstudies showed differences between classification 1 (and 1.5)point players compared to the other classifications (Molik &Kosmol, 2001; Vanlandewijck et al., 2003; Vanlandewijck,Spaepen, & Lysens, 1995). Currently, there is not a clear relation-ship between classification and mobility performance. Theimpact and content of the classification system should befurther investigated in future research.

The second hypothesis was that players competing at aninternational playing standard perform better than players at anational standard. This hypothesis proved to be true for theoverall performance time and for 12 of the 14 separate activ-ities with moderate-to-large ES (0.81–1.72). Except three activ-ities related with turn on the spot, players at an internationalstandard perform all the activities faster than national stan-dard players. The difference between national and interna-tional playing standard on the overall performance time was8.11 s, which is significantly more than the LoA calculated inthe reliability study (4.20 s). Although the findings are in linewith the hypothesis, the differences may be partly explainedby other factors than the actual capacity of the athletes inmobility performance. Possibly, due to the more professionalapproach, international players may have a more optimisedwheelchair configuration compared to national players whichmight have affected their performance on the test circuit. Theactivities, which showed no differences between playing stan-dards, were again related with turn on the spot. These activ-ities are, in addition to low reliability, not distinctive forplaying standards. Turns on the spot are frequent elementsof performance during matches and, therefore, important toinclude in the WMP test. However, time appears not to be a

reliable outcome measure for these activities. In order tooptimise the test, these activities must be further examined.At the moment, the WMP test is also analysed with data frominertial sensors using the method of Van Der Slikke, Berger,Bregman, Lagerberg, & Veeger (2015) with outcome measuressuch as velocity and acceleration.

The third hypothesis was that men perform better at theWMP test than women of the same playing standard. Except,again, for the activities related with turn on the spot, thehypothesis proved true. Men did perform all activities fasterthan women, except for the 12 m sprint with ball possession.The hypothesis is based on differences in upper body strengthand trunk stability between men and women (Gomez et al.,2014). However, for the 12 m sprint with ball possession, ball-handling skills play an important role. For the rotational move-ment combined with ball possession, the hypothesis was pro-ven. It may be possible that there is a difference in trainingfocus between the international men and women in ball hand-ling. Women may have better ball skills and with this theycompensate for their slower performance on the 12 m sprint.

In this study, three hypotheses were formulated and testedto determine the construct validity of the WMP test. Thesehypotheses are chosen based on literature and practical feasi-bility. Several other variables than classification, gender andsex could have an influence on the mobility performance.Examples of variables which may also could have been usedare floor surface and wheelchair configurations aspects suchas wheel size, camber and elbow angle. Floor surface canaffect performance due to a different rolling resistance, andthe WMP test should reveal this difference. However, for thepresent study, it was practically difficult to organise to haveplayers perform the test circuit at different floor surface. Inaddition, it should be mentioned that other variables thanmentioned in the hypothesis might have partly affected thedifferences in mobility performance. In this study, we focusedprimarily on the construct validity of the WMP test and not atvariables that best predict performance on the WMP test.

Reliability

The ICC values of the separate activities of the WMP testranged between 0.25 and 0.95, and five of the 15 outcomemeasures showed low reliability (<0.70). The ICC of four activ-ities that included a turn on the spot ranged between 0.25and 0.62. The performance time of these activities is very shortcompared to the other activities. For example, the averageduration for a turn of the spot (left) is 0.90 s with SEM of 0.1 s.The reason for these lower ICC values could be that themeasurement error of these activities is relatively high dueto the short performance times. Because of this, performancetime may not be an adequate outcome parameter in thesefour activities. In this study, the reliability between the WMPtest and retest on the 12 m sprint time was also low(ICC = 0.62). Previous research showed that time over a 15 msprint cannot be used to assess wheelchair-specific capacity(Van der Scheer, de Groot, Vegter, Veeger, & van der Woude,2014). In contrast, de Groot et al. (2012) reported a goodreliability score (ICC 0.80–0.84) for a 5-m sprint test. Thesedifferences in reliability could be explained by the differences

30 A. M. H. DE WITTE ET AL.

in handling the timing of deceleration to stop. In our study,the players had to stand still at the end of the 12 m, while inthe study of de Groot et al. (2012), the players were allowed todrive over. The potential large variation between and withinparticipants in timing of starting to decelerate and the level ofbraking (hand) forces needed to stand still at 12 m may haveresulted in a relatively large variation of performance time andthus a low reliability score. The ICC of the 12 m sprint withstops is 0.80 and well in line with the study of de Groot et al.(2012). The 12 m sprint with stops is in this case divided inthree short sprints of 3, 3 and 6 m and thus comparable indistance with the (single) 5 m in the study of de Groot et al.(2012). Although the total distance of the sprints with andwithout stops is the same, the inclusion of starts from standstill and stops seems to affect reliability. However, the designof the 12 m sprint as part of the WMP test, including theacceleration and deceleration phases, is in our opinion anessential element of mobility performance, also consideringthe results of the observations of wheelchair basketballmatches (de Witte et al., 2016).

LimitationsAll athletes performed the test in their own sports wheelchairs.Each wheelchair is individually adjusted in order to achieve anoptimal wheelchair–athlete interaction. Although wheelchairconfiguration affects mobility performance, we do not expectthis have biased our conclusions regarding validity and relia-bility of the WMP test because of the relatively large within-groups variability in wheelchair configurations. In addition, thechoice to measure wheelchair basketball players in their ownenvironment and wheelchair enhanced the external validity ofthe study. Another limitation of this study is the missing dataof activity 1 (tik-tak box) for which, in future research, thevideo set-up must be examined.

Conclusion and practical implicationsIt can be concluded that the construct validity and reliability ofthe WMP test were good for the overall performance timescore. The test can be used as a standardised mobility perfor-mance test to assess the capacity of mobility performance ofelite wheelchair athletes in wheelchair basketball. In addition,novice players might use the test to achieve a higher level ofmobility performance and monitor their progression in mobi-lity performance aspects related to elite wheelchair basketball.The overall outcome of the WMP test is reliable. However, theactivities related with turn on the spot (no. 2, 6, 10 and 14)show low reliability and construct validity.

The WMP test can be easily used to periodically monitorthe capacity of wheelchair basketball players in mobility per-formance. The test results can be used to detect strengths andweaknesses of players in different aspects of mobility perfor-mance. For example, when a player performs driving forwardactions significantly better than rotation actions – comparedwith team mates – the trainer can use these outcomes todevelop specific training schemes. In addition, the test canbe used to monitor the progress in mobility performance, todetect talented athletes and to examine whether an athlete issufficiently recovered from an injury. For research purposes,we aim to use this WMP test to examine the impact of

different wheelchair configurations on mobility performance,as recommended by Mason et al. (2013).

Disclosure statement

No potential conflict of interest was reported by the authors.

Funding

This work was supported by Taskforce for Applied Research (part ofNetherlands Organization for Scientific Research).

ORCIDAnnemarie M. H. de Witte http://orcid.org/0000-0002-5687-4351

References

Atkinson, G., & Nevill, A. M. (1998). Statistical methods for assessing measure-ment error (reliability) in variables relevant to sports medicine. SportsMedicine, 26(4), 217–238. doi:10.2165/00007256-199826040-00002

Barfield, J., & Malone, L. A. (2012). Performance test differences andparalympic team selection: Pilot study of the United States nationalwheelchair rugby team. International Journal of Sports Science &Coaching, 7(4), 715–720. doi:10.1260/1747-9541.7.4.715

Bland, J. M., & Altman, D. (1986). Statistical methods for assessing agree-ment between two methods of clinical measurement. The Lancet, 327(8476), 307–310. doi:10.1016/S0140-6736(86)90837-8

Bloxham, L. A., Bell, G. J., Bhambhani, Y., & Steadward, R. D. (2001). Timemotion analysis and physiological profile of Canadian world cup wheel-chair basketball players. Sports Medicine, Training and Rehabilitation, 10,183–198. doi:10.1080/10578310210398

Byrnes, D., & Hedrick, B. (1994). Comprehensive basketball grading system.In B. Hedrick, D. Byrnes, & L. Shaver (Eds.), Wheelchair basketball (pp.79). Washington, DC: Paralyzed Veterans of America.

Chow, J. W., Millikan, T. A., Carlton, L. G., Chae, W., Lim, Y., & Morse, M. I.(2009). Kinematic and electromyographic analysis of wheelchair propul-sion on ramps of different slopes for young men with paraplegia.Archives of Physical Medicine & Rehabilitation, 90(2), 271–278.doi:10.1016/j.apmr.2008.07.019

Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nded.). Hillsdale, NJ: Lawrence Erlbaum.

Cohen, J. (1992). Statistical power analysis. Current Directions inPsychological Science, 1(3), 98–101. doi:10.1111/cdir.1992.1.issue-3

de Groot, S., Balvers, I. J. M., Kouwenhoven, S. M., & Janssen, T. W. J. (2012).Validity and reliability of tests determining performance-related com-ponents of wheelchair basketball. Journal of Sports Sciences, 30(9), 879–887. doi:10.1080/02640414.2012.675082

de Groot, S., Valent, L. J., Fickert, R., Pluim, B., & Houdijk, H. (2016). Anincremental shuttle wheel test for wheelchair tennis players.International Journal of Sports Physiology and Performance.doi:10.1123/ijspp.2015-0598

de Witte, A. M., Hoozemans, M. J.,Berger, M. A., Veeger, D. (. E. J. )., & vander Woude, L. H. V. (2016). Do field position and playing standardinfluence athlete performance in wheelchair basketball? Journal ofSports Sciences, 34(9), 811–820. doi:10.1080/02640414.2015.1072641

Gil, S. M., Yanci, J., Otero, M., Olasagasti, J., Badiola, A., Bidaurrazaga-Letona, I., . . . Granados, C. (2015). The functional classification andfield test performance in wheelchair basketball players. Journal ofHuman Kinetics, 46(1), 219–230. doi:10.1515/hukin-2015-0050

Gomez, M. Á., Perez, J., Molik, B., Szyman, R. J., & Sampaio, J. (2014).Performance analysis of elite men’s and women’s wheelchair basketballteams. Journal of Sports Sciences, 32(11), 1066–1075. doi:10.1080/02640414.2013.879334

Goosey-Tolfrey, V., & Leicht, C. (2013). Field-based physiological testing ofwheelchair athletes. Sports Medicine, 43(2), 77–91. doi:10.1007/s40279-012-0009-6

JOURNAL OF SPORTS SCIENCES 31

Granados, C., Yanci, J., Badiola, A., Iturricastillo, A., Otero, M., Olasagasti, J., . . .Gil, S. M. (2015). Anthropometry and performance in wheelchair basket-ball. Journal of Strength & Conditioning Research (Lippincott Williams &Wilkins), 29(7), 1812–1820. doi:10.1519/JSC.0000000000000817

International Wheelchair Basketball Federation. (2014). Official wheelchair bas-ketball rules 2014. Incheon: International Wheelchair Basketball Federation.

Mason, B., Porcellato, L., van der Woude, L. H. V., & Goosey-Tolfrey, V. L.(2010). A qualitative examination of wheelchair configuration for opti-mal mobility performance in wheelchair sports: A pilot study. Journal ofRehabilitation Medicine, 42(2), 141–149. doi:10.2340/16501977-0490

Mason, B., van der Woude, L., & Goosey-Tolfrey, V. (2013). The ergonomics ofwheelchair configuration for optimal performance in the wheelchair courtsports. Sports Medicine, 43(1), 23–38. doi:10.1007/s40279-012-0005-x

Molik, B., & Kosmol, A. (2001). In search of objective criteria in wheelchairbasketball player classification. Vista, 99, 355–368.

Rhodes, J., Mason, B., Perrat, B., Smith, M., & Goosey-Tolfrey, V. (2014). Thevalidity and reliability of a novel indoor player tracking system for usewithin wheelchair court sports. Journal of Sports Sciences, 32(17), 1639–1647. doi:10.1080/02640414.2014.910608

Shrout, P. E., & Fleiss, J. L. (1979). Intraclass correlations: Uses in assessingrater reliability. Psychological Bulletin, 86(2), 420–428. doi:10.1037/0033-2909.86.2.420

Van der Scheer, J.W., de Groot, S., Vegter, R. J., Veeger, D. H., & van derWoude,L. H. V. (2014). Can a 15m-overground wheelchair sprint be used to assesswheelchair-specific anaerobic work capacity? Medical Engineering &Physics, 36(4), 432–438. doi:10.1016/j.medengphy.2014.01.003

Van der Slikke, R., Berger, M., Bregman, D., Lagerberg, A., & Veeger, H.(2015). Opportunities for measuring wheelchair kinematics in matchsettings; reliability of a three inertial sensor configuration. Journal ofBiomechanics, 48(12), 3398–3405. doi:10.1016/j.jbiomech.2015.06.001

Van der Slikke, R., Berger, M., Bregman, D., & Veeger, H. (2015). Estimatedathlete’s basketball match performance based on measurement of wheel-chair kinematics using inertial sensors. Glasgow, UK: InternationalSociety of Biomechanics.

Van der Slikke, R. M. A., Berger, M. A. M., Bregman, D. J. J., Lagerberg, A. H.,& Veeger, H. E. J. (2016). From big data to rich data, the key kinematicsof wheelchair mobility performance. Journal of Biomechanics, 49(14),3340–3346. doi:10.1016/j.jbiomech.2016.08.022

Vanlandewijck, Y., Daly, D., Spaepen, A., Theisen, D., & Pétré, L. (1999).Biomechanics in handrim wheelchair propulsion: Wheelchair-user inter-face adjustment for basketball. Education, Physical Training, Sport, 33(4),50–53.

Vanlandewijck, Y., Daly, D., & Theisen, D. (1999). Field test evaluation ofaerobic, anaerobic, and wheelchair basketball skill performances.International Journal of Sports Medicine, 20(8), 548–554. doi:10.1055/s-1999-9465

Vanlandewijck, Y., Theisen, D., & Daly, D. (2001). Wheelchair propulsionbiomechanics: Implications for wheelchair sports. Sports Medicine, 31(5),339–367. doi:10.2165/00007256-200131050-00005

Vanlandewijck, Y. C., Evaggelinou, C., Daly, D. D., Van Houtte, S., Verellen,J., Aspeslagh, V., . . . Zwakhoven, B. (2003). Proportionality in wheelchairbasketball classification. Adapted Physical Activity Quarterly, 20(4), 369–380. doi:10.1123/apaq.20.4.369

Vanlandewijck, Y. C., Spaepen, A. J., & Lysens, R. J. (1995). Relationshipbetween the level of physical impairment and sports performance inelite wheelchair. Adapted Physical Activity Quarterly, 12(2), 139–150.doi:10.1123/apaq.12.2.139

Yilla, A. B., & Sherrill, C. (1998). Validating the beck battery of quad rugbyskill tests. Adapted Physical Activity Quarterly, 15, 155–167. doi:10.1123/apaq.15.2.155

32 A. M. H. DE WITTE ET AL.